TitanbaseDocs

Local-first workflow

Titanbase is local-first. The editor runs in your browser, your schema lives as a file on your disk, and exports happen locally. No account is required, and the local product does not need the cloud.

Note

Your schema stays local. Titanbase does not upload your .titan.json, and it does not connect to any live database.

What "local-first" means here#

  • No account required — open the editor and start designing.
  • Your schema stays local — it is saved as a .titan.json file you control.
  • Works without the cloud — the editor, .titan.json, validation, and exporters all run locally.
  • No live database connection — Titanbase designs schemas; it does not read from or write to a running database.

The everyday loop#

  1. Design — create tables, columns, relations, indexes, and enums in the editor.
  2. Validate — review local diagnostics and fix any errors.
  3. Export — generate developer-ready PostgreSQL, Mermaid, Prisma, or Drizzle output.
  4. Save — write a portable .titan.json and commit it to version control.

Your file is the source of truth#

.titan.json is plain JSON. Because it is a file you own:

  • you can keep it in git and review changes in pull requests;
  • you can reopen it in the editor anytime;
  • you are never locked into a single tool or service.

The editor stores visual-only data (table positions) under metadata.editor, separate from the schema's meaning. See the .titan.json format reference for the full structure.

Why no cloud is required#

The schema engine (@titanbase/core), the editor, and the exporters are all client-side. Designing, validating, and exporting never leave your machine.

Future: an optional cloud layer#

Status: Planned

Titanbase Cloud may become a future, optional product layer for teams. It would not be required to use the local editor, .titan.json, exporters, or the planned desktop app and CLI. See Titanbase Cloud for what is being considered.